Skip to content


Our Goal

AI Playtesting aims to create an AI that can playtest combat based PVE card games and help balance it. We expect that the AI will identify dominant strategies or unbalanced cards and generate playtest data to help game designers to do the balancing. To do this we choose Slay the Spire as our target game and we are developing our own machine learning algorithm. We will document our findings and experiment through the semester and publish it afterwards.

Our Method

We created a tool for designers to balance a game called Slay the Spire.

  • Design : build your deck by adding Slay the Spire cards, or even customize your cards.
  • Training : train the AI agent using Deep Reinforcement Learning algorithms.
  • Playtest : let AI playtest and display statistics for identifying imbalance and updating the rules of the game.